##### SST comparison with in situ observations #####

The evaluation is performed for the SST product of the satellite sensors. 
These satellite data are compared to IQUAM V2(in situ SST quality monitor version 2) in situ buoy data.
IQUAM in situ observations are downloaded from IQUAM homepage (http://www.star.nesdis.noaa.gov/sod/sst/iquam/v2/).
We calculate bias and RMSE (Root Mean Square Error) after making in situ to satellite match-up data.

How to make match-up data is 

To perform SST comparison, buoy observations are collocated with the 10 satellite observations within 30 kilometers and 2 hours.

To calculate the bias and RMSE, satellite observations are collocated under the condition below. 
a. At least 2 footprints are available with good quality.
b. Maximum difference among 10 AMSR2 footprints is less than 3 degree.

If these conditions are satisfied, we calculate the bias and RMSE between buoy and average of the nearest 10 available satellite.


How to calculate bias is 
    diff(i)=satellite(i)-insitu(i)
    bias=(diff(1)+.......diff(N-1)+diff(N))/N
How to calculate RMSE is  
    RMSE*RMSE=((diff(1)-bias)*(diff(1)-bias)+.......+(diff(N)-bias)*(diff(N)-bias))/(N-1)

The left panels let you choose which sensor and product to check. 
At last if you click Display button, bias and RMSE are plotted for each area.

AREA definition is 
 : 0      All Ocean(-90<=lat<90 0<=lon<360)
 : 1      The Indian Ocean(-40<=lat<40 30<=lon<100)
 : 2      The South Atlantic(-40<=lat<10 -70<=lon<30)
 : 3      The North Atlantic(10<=lat<40 -10<=lon<30)
 : 4      The South-West Pacific(-40<=lat<10 100<=lon<180)
 : 5      The North-West Pacific(10<=lat<40  100<=lon<180)
 : 6      The South-East Pacific(-40<=lat<10 -180<=lon<-70)
 : 7      The North-East Pacific(10<=lat<40 -180<=lon<-100)
 : 8      The North Pole(40<=lat<90 -180<=lon<180)
 : 9      The South Pole (-90<=lat<-40 -180<=lon<180)

The upper left map shows large bias points.
Red : AMSR2 SST - Buoy SST >=  3 degC
Blue: AMSR2 SST - Buoy SST <= -3 degC
Please click ">Point Image" to see large bias points in each month.